? Totally Off Topic: Pandora |
| Top 40 searches including the word blog ?
August 31, 2005
Market intelligence corrupted by spam
Blogs will fail to clearly reflect the world's ongoing conversation, as promised, until a solution is found to weed out the spam. This is a problem for companies monitoring blogs for market intelligence. Example: I was on a conference call a couple days ago with executives at Factiva, the Dow Jones-Reuters joint venture. They were showing me their new Factive Insight reputation intelligence tool. In a test case, we looked at press and blog coverage from Hewlett-Packard's perspective. We saw all kinds of charts showing how much was being written about printers, the stock price, the new CEO. But when we clicked to see what a specific blog was saying, we came face-to-face with a spam blog. It was an embarrassing moment for Factiva. But until spam is under control, they'e hardly the only ones in the blog world with corrupted data.
I'll be out all day on interviews. Back tomorrow.
spam and other abuses
TrackBack URL for this entry:
Listed below are links to weblogs that reference Market intelligence corrupted by spam:
? Filtering Spam from Blog Analysis from Cymfony's Marketing Insight
Today Stephan Baker of Business Week mentioned a major problem that most people don?? realize can corrupt blog analysis. SPAM. Keeping spam out of analytics is one of the most important things we do for our clients. It?? the old [Read More]
Tracked on August 31, 2005 12:13 PM
I have not looked at Factiva. It sounds like they have a reputation problem or something. Joint ventures don't always work. Maybe they will figure it out.
Posted by: Jim Dermitt at August 31, 2005 08:29 AM
Spam blogs or splogs are definitely a problem and it sounds like in this instance it caused a little bit of embarrassment. Splog Reporter is trying to address the spam blog or splog problem that is plaguing the blogosphere. We recently released a version 2 of Splog Reporter with added features to make it easier for users to report spam blogs or splogs. Some of the key new features include a bookmarklets as well as a Firefox extension that makes reporting splog nearly a one click process. For more information go to: http://www.splogreporter.com
Posted by: Frank Gruber at August 31, 2005 11:48 AM
I imagine that situation was embarassing but I've never encountered an automated system filtering large amounts of data that wasn't prone to such difficulties. The reality is that blog spam won't be stopped and tech solutions can only do so much. Actual human evaluation of data is what's missing from these systems.
Posted by: Clyde Smith at September 1, 2005 04:39 AM
PS - I do think that automated systems work ok when the sources that are being spidered are screened by humans. That's part of why Google News does a reasonable job. They screen the sources that are then tracked automatically.
Posted by: Clyde Smith at September 1, 2005 04:43 AM
Stephen, you raised a great point on spam blogs. The issue of spam blogs challenges the entire blog-monitoring industry, and it's one that all major players are working to manage. The issue is complicated because, on the one hand, executives tracking information that may impact their company?? reputation may want to exclude spam blogs from the content sets they monitor. However, some may want to monitor all media??including spam blogs, as it is critical to know the entirety of what?? being said, regardless of credibility. It?? up to the individual.
While the relative volume of spam blogs found in the market intelligence gleaned from Factiva Insight: Reputation Intelligence is small, Factiva does want to address what the company is doing to give our clients the option to reduce the occurrence of spam blogs in the product?? content collection.
?? In all of our customer engagements, Factiva's text mining depends significantly on editors crafting and refining client-specific text-mining algorithms. Upon seeing initial results including spam blogs, the editor could refine the query to ignore those types of posts.
?? As additional posts from spam blogs are found in Factiva's archive, Factiva editors will have the option to ignore them in client analyses.
The blog content in the Factiva Insight: Reputation Intelligence system is aggregated by our partner, Intelliseek. Following are some actions that Intelliseek has told Factiva it is taking to help reduce spam blogs from their feeds:
?? A few months ago Intelliseek created a task force to arrive at a comprehensive approach for reducing spam in their feeds.
?? In early September they will implement changes, based on their research, which will start to reduce the flow of spam.
?? By the end of October, a more significant reduction is expected as more of the changes come online.
?? Additional steps are being taken to remove already indexed spam from their archive.
Alan Scott, CMO, Factiva
Posted by: Alan Scott at September 6, 2005 04:11 PM